| Introduction |
Appendices |
||
| Legal Cases |
|||
Chapter 5: Doubt and its incorporation into medical
decision making
In medical discussions concerning PVS, a certain cluster of propositions is common:
|
P1 - |
‘Patient X is in a PVS.’ |
|
P2 - |
‘Patient X manifests only reflex actions.’ |
|
P3 - |
‘Patient X has no consciousness.’ |
|
P4 - |
‘Patient X cannot feel pain.’ |
|
P5 - |
‘Patient X is no longer a ‘person’.’ [i]
|
These propositions are not regarded as being independent propositions, but as being part of a deductive scheme with P1 implying P2; P2 implying P3 and P5, and P3 implying P4 and P5.
It is possible to critically analyse such a proposition cluster on two levels - that of the truth of the individual propositions, or that of the validity of the inferences between the propositions. On the level of the individual propositions; criticism may take the form of an assertion that the individual propositions are not well founded in that they are not capable of verification - i.e. there is no clear, unambiguous, procedure for determining, in any particular case, whether or not the propositions are true. On the level of cluster as a whole; criticism may take the form of an assertion that the deductive links between the propositions are not valid - e.g. that P2 does not imply P3 or that P3 does not imply either P4.[ii]
This analysis has, in essence, been the task undertaken in the last four chapters. The reason for such an undertaking was to show that, in relation to each of the propositions and inferences, a doubt, or uncertainty, existed. The magnitude of the doubt differed as between the propositions, but it was not negligible; for example, in relation to P1 it was of the order of 50%.[iii]
An attempt at such a rigorous critical analysis seems, at first sight, somehow inappropriate when applied to the propositions of clinical medicine. However, as medicine often clothes itself in the mantle of science, and an analysis, such as is proposed above, would be a commonplace in a discussion of physics or biology, it follows that the sense of inappropriateness does not arise from the propositions in so far as they are scientific propositions. The propositions P1, P2 , P3 and P4 are somehow more than purely propositions of science and the inappropriateness flows from this ‘more’.[iv] What could this ‘more’ be? This question is explored in Section 1, where it is suggested that the answer to the question lies in the fact that the propositions of clinical medicine - unlike those of pure science - have a normative aspect.
In any clinical situation the basic issues - such as the diagnosis, the appropriate treatment to be adopted, the side-effects of such treatment, the prognosis - are to some extent, open to doubt. The magnitude of the doubt may, of course, vary but it seems intrinsic to the clinical situation. How should such doubt be incorporated into the making of clinical decisions? Should, for example, decisions be based on a simple balance of probabilities basis? Such an approach would dictate that the most probable diagnosis should be chosen and that then the most effective treatment - judged on a probability basis - should be performed. However, consider the following example: a patient presents with a severe leg infection; the diagnosing physician believes that there is a 51% chance that it is gangrene; a 49% chance that it is a rare fungal infection. The best treatment for gangrene is amputation of the leg, that for the fungal infection is application of an ointment. A ‘decision procedure’ based on choosing that which was most probable would result in the amputation of the leg. This suggests that, whereas to choose the most probable diagnosis is, in many cases, the correct strategy, it should not necessarily be chosen if the unwanted effects of an incorrect choice, are radically disproportionate. An analysis of such decision procedures is developed in Section 2. It is applied to problems concerning the treatment of PVS patients in Section 3. Section 4 draws some conclusions.
Section 1: The normative aspects of medical propositions.
The propositions ‘water is composed of atoms of hydrogen and oxygen’ or ‘force equals mass times acceleration’ carry of themselves no moral imperative, no urge to action. Contrast these with the proposition ‘Mr. Smith is suffering from appendicitis’. This latter proposition is, in one sense, a scientific proposition - the terms are unambiguous and the assertion is open to verification - but it is not just a proposition of chemistry or biology. Such a medical proposition is normative in that it implies an ethical obligation on certain people to act in a particular fashion; to say that a patient is suffering from appendicitis carries the moral obligation on any medical person, who has a duty of care to that patient,[v] to act in a medically appropriate fashion.
Thus, medical propositions have both a normative and a scientific content.[vi] The canons of criticism which are considered appropriate for scientific propositions are different from those considered appropriate for normative propositions. Herein lies the reason for the seeming inappropriateness of the critique of propositions P1, P2 , P3 and P4 discussed earlier - for this was a critique which regarded the propositions as purely scientific and disregarded their normative content.
The following example clearly shows the difference between interpreting a particular proposition in a purely scientific manner, and interpreting it normatively; it also shows that the accepted mode of criticism differs depending on which interpretation is active. Imagine strolling beside the sea and suddenly hearing a shout “That boy is drowning!”. A purely scientific attitude to this proposition would seek to determine its truthfulness - ‘Is the boy really drowning? Perhaps we should wait and see!’; and a scientific criticism of the proposition would be that a doubt existed that he was drowning, and that a replication of the event was necessary. Such a response would rightly be regarded as inhuman and monstrously inappropriate because it would be oblivious to the normative content of the proposition - the implication that one should attempt to rescue the boy. An example of criticism from a normative perspective would be that the bystander did not attempt to rescue the boy; a reply from the bystander that he wanted to be absolutely certain that the boy was drowning and that he was attempting to resolve all these doubts before acting, would be considered inappropriate. In such circumstances a commonly accepted ethical perspective would be that, although one might have doubts as to whether the boy was drowning, if, on the balance of probabilities one believed that he was drowning, then one should put these doubts aside and take whatever measures one could to rescue him.
This suggests that the attitude to doubt and its resolution, is an important manifestations of the difference between a set of propositions being interpreted from a purely scientific standpoint and from a normative standpoint. Should doubts be actively sought out? Should a distinction be drawn between ‘major’ doubts - those likely to influence a judgement made on the balance of probability - and others? or, as in law, between reasonable doubts and others? Is there some proportionality between the magnitude of a doubt and its ethical relevance? Answers to such questions are to a large measure determined by the standpoint adopted.
The role of doubt in scientific, normative and medical discourses contrasted.
In scientific discourse
In a purely scientific discourse the possibility of doubting a particular proposition is the generator for future research and scientific development; doubt is to be cultivated; it is a boon, the seed of growth. If it does not seem an inappropriate use of language, one might say that within a scientific discourse there is a moral obligation to unearth doubts. Moreover, it is the possibility of doubt rather than the magnitude of the doubt that is of importance: if a result can be experimentally verified in all but one case, then what matters is that the result is not universally valid and not that the occurrence is once in a hundred cases rather than once in a billion. The situation is otherwise in a normative discourse.
In normative discourse
A normative proposition has the form ‘If condition X occurs, then Y must be done.’ There are two obligations implicit in this, the obligation to determine whether, in fact, condition X has actually occurred and, if it has, the obligation to do Y. The first obligation is similar to the obligation of the scientist to determine the truth of the proposition ‘X has occurred’ but this similarity is superficial. There may indeed be some circumstances where the obligation to determine whether X has occurred is an absolute one - i.e. such that under no circumstances must Y be done unless X has occurred; an example from clinical medicine would be the determination of death before the signing of the death certificate; but generally not only is the obligation to determine the condition X not only not absolute, but there is a time constraint, so that an undue time spent on the determination of X is at the cost of time that should be spent on Y. To spend time achieving certainty in determining X, resulting in time not remaining to do Y, would indeed be considered to be an abject failure. A corollary of this is that under conditions of urgency there may well be a moral obligation not to raise minor doubts as to the truth of X, because to do so would mitigate against the possibility of Y being accomplished; this is particularly so if Y - as in the case of health interventions - is regarded as being generally salutary.
In medical discourse
The situation is similar in clinical
medicine. There the normative aspect of
the situation - the obligation to ease the patient’s suffering - is ever present and ‘doubt’ has a much
more limited role than in a scientific investigation. A tension exists between the normative and the scientific aspects
of the medical discourse; and a concentration on the scientific aspects of a
problem - i.e. a concern with
resolving uncertainties, or doubts - may be seen as being incompatible with, or
destructive of, the normative aspects.
The diagnosis, and the decisions that flow from it, must often be made
under conditions of urgency and though major doubts cannot be ignored, fanciful
or ‘academic’ doubts must be set aside because, if they are entertained, they
not only create indecision - which can cripple effective action - but they
require time to resolve and such time may well be of the essence: the belief
being that the patient may well be dead before all the doubts as to their
illness can be resolved.[vii] This tendency to dismiss doubts may be
reinforced by the generally accepted perception that it is better to err on the
side of a health intervention than not - the belief that a health intervention,
when not actually required, is less damaging than an absence of intervention
when it is required, gives added momentum to the tendency to disregard minor
doubts. Thus, in clinical medicine, in
contrast to pure science, the moral
obligation is understood to imply that doubts should be minimised rather than
nurtured.
Hence the reason for the seeming inappropriateness of the analysis discussed earlier[viii] lies in the fact that - in seeking to establish lacunae where doubt could gain a foothold - it was embarking on a project whose value is not necessarily accepted in clinical medicine.
The compartmentalising of medical problems
and its effect on ethical judgements
The exigencies of the clinical situation may demand not only that doubts, other than major doubts, be disregarded but also that different aspects of a presenting situation be separated. A medical professional, in order not to be overwhelmed by the problem with which he is faced, must compartmentalise it. He must split the problem into simpler problems and solve these separately. First, he must diagnose: he must ask ‘What is the most probable medical condition obtaining at present?’; then he must put aside any uncertainty as to the diagnosis - as otherwise his ability to act effectively is hampered - and he must ask ‘Given the condition as diagnosed, what is the prognosis of this condition and what is the most effective treatment?’
These two aspects of clinical medicine - the propensity to disregard ‘non-major’ doubts and the compartmentalisation of a problem into sub-problems - whilst often necessary within the confines within which clinical medicine must operate, can, if carried over into settings other than those requiring emergency medical intervention, have profoundly unethical and inappropriate consequences; this is particularly so in relation ethical discussions of medical issues. Simplifications which may well be justified in a clinical situation have no justification in the more relaxed conditions within which medical ethicists debate.[ix]
Conclusion 5 -1 : ‘Doubt’ plays an unequivocally positive role in a scientific discourse in that it is the very seed for future development. In contrast, its role in a medical discourse is ambivalent: from a scientific perspective, ‘doubt’ is valued; however, from a clinical perspective, the cultivation of doubt may be seen as destructive of appropriate, and timely, intervention. Medical ethics appears to assign a role to ‘doubt’ similar to that accorded to it by clinical medicine without, however, having the justifications for doing so which are available to clinical medicine.
Section 2: A schema to permit the incorporation of uncertainty into the making of medical decisions.
In the face of uncertainty as to which of two particular states of affairs persists, it may be thought that the correct course is to assume to be true, that which is most probably true [‘Principle of choosing the most probable’ ]. A moment’s reflection will show that this principle, when used as a decision procedure for making ethical choices, is not appropriate because it does not seek to minimise the evils or maximise the goods that would flow from a choice between two possibilities; an example - which I have called the ‘rescue problem’ - shows this to be so.
The rescue problem
Let us assume that after an earthquake some people have been trapped in the ruins of a building and, because of extreme weather conditions, it is most improbable that any have survived; rescue operations are difficult and costly. Should the rescuers act on the most probable hypothesis? Or should they choose some alternative principle? The various options and eventualities are set out in Table 5-1:
|
line |
Rescuer’s option |
Actual situation
(unknown) |
Ethically
relevant consequences |
|
1. |
A: Act as if survivors |
There are
survivors |
In hindsight
this was the appropriate action |
|
2. |
(ditto) |
There are no
survivors |
Economic cost
but no loss of human life |
|
3. |
B: Act as if no survivors |
There are
survivors |
Saving of
economic cost, loss of human life |
|
4. |
(ditto) |
There are no
survivors |
In hindsight
this was the appropriate action |
Table 5-1: The Rescue Problem
Using the ‘principle of choosing the most probable’ would suggest that option B should be chosen. This is so irrespective of the economic costs of attempting a rescue. If the situation resolves according to line 4 then the most appropriate action will, fortuitously, have been chosen. If, however, it resolves according to line 3 then the consequences are tragic; this is particularly so if the economic cost of rescuing the survivors was minimal. It is clear that a more sophisticated decision procedure is required, one which incorporates the consequences of the choices, rather than simply taking note of the probabilities of the various choices being true. The ethical difficulty is caused - not by lines 1 or 4 - but by lines 2 and 3. So the focus must be placed on the unwanted consequences of choices which are, in hindsight, incorrect. But the choice cannot be made by simply comparing the respective unwanted outcomes for this takes no account of the likelihood of the initial probabilities. What is needed is a method of bringing together both the probability of a state of affairs, and the consequences of an incorrect judgement, into one index so that a more rational choice can be made. This involves estimating, or quantifying, the unwanted consequences[x] and then multiplying each of these estimates by their respective probabilities - in effect this is a method of estimating the likelihood of the unwanted outcomes corresponding to the differing choices - and then choosing the option corresponding to the minimum likelihood. An example (set out in Table 5-2) may help clarify. In order that the underlying method be clearly explained it has been assumed that the probability of survivors being present is 1 in 10. It is also assumed that the unwanted consequences of a wrong decision can be quantified and are in the ratio 1:1000, i.e.:
|
Consequences of assuming that there were no
survivors whilst there were, in fact, survivors |
= |
1000 |
|
Consequences of assuming that there were survivors
whilst there were, in fact, no survivors |
|
1 |
|
line |
Rescuers have option |
Actual situation (unknown) |
P (Probabilities) |
Ethically relevant, consequences of an incorrect
decision. |
C* |
Product of P and C |
|
2. |
A: act as if survivors |
There are no
survivors |
9/10
(Probability of no survivors) |
Considerable
economic cost but no loss of human life |
1 |
(9/10)x1 = 0.9 |
|
3. |
B: act as if no survivors |
There are
survivors |
1/10 (Probability of
survivors) |
Saving of
economic cost, loss of human life |
1000 |
(1/10)x1000 = 100 |
*relevant consequences of an incorrect decision, quantified and
expressed as a ratio.
Table 5-2: The rescue Problem - analysis of unwanted consequences
The last column in Table 5-2 attempts to quantify the likelihood of the unwanted consequences of each course of action if, in hindsight, a factually incorrect choice was made. If the ethical principle adopted is that one should minimise the evils (the unwanted consequences) of one’s actions, then one should minimise the index (P multiplied by C). In the example given, since 100 is far in excess of 0.9 one should act as if there are survivors; this is so even though it is not the most probable situation.
Conclusion 5 -2 : Acting on the basis that those circumstances which are most probably true, are true, does not ensure that the unwished for consequences that flow from a (then unknowably) incorrect choice are minimised. In such cases an index should be compiled of the probabilities of various eventualities, weighted in proportion to the magnitude of their respective unwanted consequences, and the eventuality be chosen which corresponds to the minimum index. In short, the decision procedure to be adopted in such cases of incomplete knowledge, is that one should act so that the unwished-for consequences that flow from a (then unknowably) incorrect choice are minimised.
Section 3: Application of the decision schema to the PVS situation.
Let us first consider the traditional ethical approach to PVS/pain/consciousness judgements. A patient presents with the symptoms of PVS and, after a thorough examination, is diagnosed as such. Once this diagnosis is made then the judgement that the patient has no consciousness or cannot experience pain, is said to follow either as a matter of strictly logical necessity or else to be of such high probability that to suggest otherwise is to raise purely fanciful doubts. The suggestion that the patient be treated as if they are conscious and can experience pain elicits the immediate reply that to do so would be to act contrary to what is known to be true and that to follow such a course of action could hardly be ethical; and, furthermore, that an acceptance of the ability of the patient to experience pain would be seen as destructive of the diagnosis of PVS. I suggest that this bifurcation of the problem into diagnosis and treatment, with the doubts being separately resolved on the basis of accepting that which is most probable, distorts PVS/pain/consciousness judgements. Let us apply the analysis undertaken in the rescue problem to these judgements.
The rescue problem applied to the PVS situation.
To apply the analysis of the rescue problem to the situation of PVS patients and the possibility of their experiencing pain, it is first necessary to estimate the probability of a PVS-like patient being diagnosed, wrongly, as being ‘unable to experience pain’. At first sight it may appear fanciful to attempt to ascribe probabilities to such a proposition. It is not: if, as is usual, the proposition that ‘a PVS patient cannot experience pain’ is held to follow as a matter of strict deduction from the diagnosis of PVS then an estimate of the rate of misdiagnosis of PVS would, subject to certain reservations, appear to give a minimum estimate of probability. However, the problem is more complicated than being simply one of misdiagnosis. Three, intertwined, issues are involved:
(i) the possibility that a PVS-like patient be misdiagnosed as being PVS.
(ii) the possibility that a misdiagnosed PVS patient has experienced pain whilst diagnosed as PVS.
(iii) The possibility that a correctly diagnosed PVS patient has experienced pain whilst diagnosed as PVS.
We have seen[xi] that the best estimate for the rate of misdiagnosis of PVS is of the order of 50%. This suggests that even if the ‘principle of choosing the most probable’ is adopted, PVS patients should be treated as being able to experience pain. This, however, presumes that in each case of misdiagnosis the patient was - whilst diagnosed as PVS - at some stage able to experience pain. As this inference may not be fully justified let us assume, for the sake of illustration, that 30% of patients diagnosed as PVS are both wrongly diagnosed and can at some stage (whilst diagnosed as PVS) experience pain. The possibility that a correctly diagnosed PVS patient has experienced pain[xii] is more problematic because it directly raises the problem of the definition of PVS. This has been discussed exhaustively in Chapter 4 and, for the purposes of this discussion, it is clear that if PVS is defined by the absence of behavioural manifestations of consciousness then there is a clear possibility that a correctly diagnosed PVS patient could experience pain. The inclusion of such patients would obviously increase the estimate derived from misdiagnosis studies; nonetheless let us assume that the conservative estimate of 30% still applies.
We have next to make an estimate of the detrimental consequences of treating an aware patient as not aware, in comparison to the detrimental consequences of treating an unaware patient as aware. The various consequences are set out in Table 5-3:
|
line |
Option: Patient treated as
if aware and able to feel pain |
Patient in fact aware and
able to feel pain |
Ethical consequences |
|
|
1. |
A: Yes |
Yes |
Considerable
positive consequences to all concerned |
|
|
2. |
A: Yes |
No |
Professional carers: |
slight
economical cost |
|
|
|
|
Patient’s family: |
possible benefits |
|
|
|
|
Patient: |
no detrimental
effect |
|
3. |
B: No |
Yes |
Professional carers: |
slight
economical benefits |
|
|
|
|
Patient’s family: |
possible
negative consequences |
|
|
|
|
Patient: |
very severe
negative consequences |
|
4. |
B: No |
No |
Slight ethical
consequences; possible detriment to the family |
|
Table 5-3: Analysis of doubt applied to the PVS Problem.
As in the rescue problem, the analysis must concentrate on comparing the unwanted consequences of, in hindsight, incorrect choices i.e. lines 2 and 3. We must compare the consequences of treating a patient as unable to experience pain (when they are able), with the consequence of treating one as able to experience pain (when they unable). In essence, this reduces to comparing the suffering of a patient who is untreated for pain, with the cost of providing analgesic procedures when they are, in fact, unnecessary. It is invidious to make such comparisons but they are required for the analysis and I have chosen a ratio of 1000:1 for illustrative purposes.
|
(i) |
(ii) |
(iii) |
(iv) |
(v) |
|
|
|
Treat as able to experience pain |
In fact able to experience pain |
Probability (P) that patient able to experience
pain - 30% |
The unwanted results expressed as a ratio (R) 1000:1 |
Product of P and R |
|
2. |
A: Yes |
No |
0.7 |
1 |
0.7 |
|
3. |
B: No |
Yes |
0.3 |
1000 |
300 |
Table
5-4: The PVS Problem: a tentative attempt at quantification for the purposes of
illustration
Because 300 is greater than 0.7 the analysis as developed in the rescue problem, implies that such patients should be treated as if they can experience pain. By asking what figures are required in either Columns (iii) or (iv) to make the indices in column (v) equal, the conclusion can be restated in alternative ways:
Conclusion 5 -3 : If it is accepted that the probability of a patient diagnosed as PVS being able to experience pain is 30%; then, in order to justify analgesic treatment being withheld from PVS patients, it must be accepted that the non-treatment of PVS patients for pain, when they can, in fact, experience pain, is no more than just over twice as abhorrent as the treatment of PVS patients as being able to experience pain, when, in fact, they cannot.
Conclusion 5 -4 : If it is accepted that the non-treatment of PVS patients for pain, when they can, in fact, experience pain, is of the order of 1000 times as abhorrent as the treatment of PVS patients as being able to experience pain when, in fact, they cannot; then, in order to justify analgesic treatment being withheld from PVS patients, it must be accepted that the possibility of a PVS patients being able to experience pain is of the order of 1 case in every 1000.
A note on the phrase ‘as if’
The phrase ‘as if’ has been used in the context of suggesting that PVS patients should be treated as if they are conscious. The use of this phrase is not meant to suggest that some pretence be enacted but rather that an attitude towards PVS patients be cultivated which acknowledges that some patients who are diagnosed as PVS undoubtedly are conscious and able to experience pain but that we have no way of identifying which patients are, in fact, in pain.
A not dissimilar problem relating to the attitude to be adopted by medical staff in situations where there is incomplete knowledge occurs in relation to the disposal of used syringes, particularly in so far as they present a danger of contracting aids or hepatitis. The attitude of medical staff in such cases - in that they presume some are contaminated but they do not know which - exemplify the sense of ‘as if ‘ that is being proposed: the medical staff treat all syringes as if they were contaminated.
Section 4: Conclusions
The conclusion to be drawn from this chapter is that - especially in view of the extent of misdiagnosis of PVS:
Conclusion 5 -5 : All patients diagnosed as PVS, should be treated as if they are conscious and can experience pain.
It could be argued that erroneously treating a patient as being able to feel pain might cause distress to the patient’s family; however, this would be a misunderstanding of what is being suggested which is not that the patient can feel pain, but rather that there is uncertainty, and in the face of this uncertainty the ethical practice should be to treat the patient as if they can feel such pain. It does not mean that the withdrawal of life-sustaining treatment or the withdrawal of ANH should not occur;[xiii] other than to ensure that pain relief is given, it is simply not relevant to these decisions.[xiv]
The conclusions that were established in this Chapter are:
Conclusion 5 -1 : ‘Doubt’ plays an unequivocally positive role in a scientific discourse in that it is the very seed for future development. In contrast, its role in a medical discourse is ambivalent: from a scientific perspective, ‘doubt’ is valued; however, from a clinical perspective, the cultivation of doubt may be seen as destructive of appropriate, and timely, intervention. Medical ethics appears to assign a role to ‘doubt’ similar to that accorded to it by clinical medicine without, however, having the justifications for doing so which are available to clinical medicine.
Conclusion 5 -2 : Acting on the basis that those circumstances which are most probably true, are true, does not ensure that the unwished for consequences that flow from a (then unknowably) incorrect choice are minimised. In such cases an index should be compiled of the probabilities of various eventualities, weighted in proportion to the magnitude of their respective unwanted consequences, and the eventuality be chosen which corresponds to the minimum index. In short, the decision procedure to be adopted in such cases of incomplete knowledge, is that one should act so that the unwished-for consequences that flow from a (then unknowably) incorrect choice are minimised.
Conclusion 5 -3 : If it is accepted that the probability of a patient diagnosed as PVS being able to experience pain is 30%; then, in order to justify analgesic treatment being withheld from PVS patients, it must be accepted that the non-treatment of PVS patients for pain, when they can, in fact, experience pain, is no more than just over twice as abhorrent as the treatment of PVS patients as being able to experience pain, when, in fact, they cannot.
Conclusion 5 -4 : If it is accepted that the non-treatment of PVS patients for pain, when they can, in fact, experience pain, is of the order of 1000 times as abhorrent as the treatment of PVS patients as being able to experience pain when, in fact, they cannot; then, in order to justify analgesic treatment being withheld from PVS patients, it must be accepted that the possibility of a PVS patients being able to experience pain is of the order of 1 case in every 1000.
Conclusion 5 -5 : All patients diagnosed as PVS, should be treated as if they are conscious and can experience pain.
[i] It may be countered that the concept of ‘personhood’ plays no role, and should play no role, in clinical medicine. However, whilst it may not feature explicitly in medical discourse, implicitly it plays a pivotal role in discussions of both infant disability, and treatment withdrawal. Its role in, for example, discussions of anencephaly is no less important for being unstated; indeed the fact that its role is unstated implies that it is unexamined and thus is of greater importance - in that its use is unfettered - than if it was made explicit. As an example of its implicit use in discussions concerning PVS patients, I cite a passage from an article by Raanan Gillon:
“... Dr. Andrews cites cases of recovery from the persistent vegetative state. These include a patient who, after three years in the persistent vegetative state, recovered sufficient consciousness to smile at cartoons, to show pleasure when his wife was present, and to show distress when she was absent. If resources were unlimited ... then the treatment should continue. But resources are severely limited ...”
[Raanan Gillon. Patients in the persistent vegetative state: a response to Dr Andrews. British Medical Journal, (1993), p.1603]
[ii] As discussed in Chapter 4, PVS is often defined to be such that P1 implies P3, although more circumspect formulations of the definition - e.g. Jennett and Plum (1972) - assert only that P1 implies P2. The deductive link between P1 and P3 ,or P1 and P2 as the case may be, is thus not open to question.
[iii] Conclusion 4 -12 The
best estimate for the rate of misdiagnosis of PVS is of the order of 50%. This
implies that the diagnosis of PVS is essentially a random process.
[iv] P5 - in that it is obviously not a purely scientific proposition - is not included in this discussion.
[v] such as the ‘patient’s doctor’.
[vi] The balance between these constituents varies depending on the circumstances: a medical diagnosis based on observations of a patient who appears to be in immanent danger of death has far greater normative content than a medical diagnosis based on the findings of an autopsy.
[vii] A perception echoed by the following Buddhist parable:
"It's just as if a man were wounded with an arrow thickly smeared with poison. His friends and companions, kinsmen and relatives would provide him with a surgeon, and the man would say, 'I won't have this arrow removed until I know whether the man who wounded me was a noble warrior, a priest, a merchant, or a worker.' He would say, 'I won't have this arrow removed until I know the given name and clan name of the man who wounded me ... until I know whether he was tall, medium, or short ... until I know whether he was dark, ruddy-brown, or golden-coloured ... until I know his home village, town, or city ... until I know whether the bow with which I was wounded was a long bow or a crossbow ... until I know whether the bowstring with which I was wounded was fibre, bamboo threads, sinew, hemp, or bark ... until I know whether the shaft with which I was wounded was wild or cultivated ... until I know whether the feathers of the shaft with which I was wounded were those of a vulture, a stork, a hawk, a peacock, or another bird ... until I know whether the shaft with which I was wounded was bound with the sinew of an ox, a water buffalo, a languor, or a monkey’. He would say, 'I won't have this arrow removed until I know whether the shaft with which I was wounded was that of a common arrow, a curved arrow, a barbed, a calf-toothed, or an oleander arrow.' The man would die and those things would still remain unknown to him.”
[From ‘The Cula-Malunkyovada Sutta’ (MN 63, The Shorter Instructions to Malunkya), included in the Middle-length Discourses of the Buddha; Internet source: http://www.accesstoinsight.org/canon/majjhima/mn63.html ]
[viii] i.e. of P1, P2, P3, P4, and their interrelationships.
[ix] I refer to, for example, Ronald Dworkin’s description of PVS patients as “... permanently damaged in a way that rules out any return to consciousness. They are capable of no sensations and no thoughts.” (quoted earlier - see Chapter 4, Section 4).
Borthwick [Borthwick (1995b) p.208] concisely states the position being advocated:
“If the discipline of medical ethics
cannot cope with uncertainty then it is useless in the real world.”
I suggest, however, that he understates the position: not only is it useless, but it is positively harmful.
[x] At least in the form of a ratio so that, for example, consequence A is 1000 times worse that consequence B.
The following betting example shows the point at issue more clearly because the consequences are already quantified:
a man is betting on whether a card taken from a deck is an ace or a spade. The probabilities are as follows:
|
choose ace |
4/52 = 0.077 |
|
choose spade |
13/52 = 0.25 |
He is then told that if he chooses an ace and is correct he wins £1000, but that if he is incorrect he must pay £1000; if he chooses a spade and is correct he wins £300, but if he is incorrect he must pay £300. What should he do to minimise his likely losses? What should he do to maximise his likely gains?
By multiplying each consequence by its associate probability he estimates the likely consequence of each course of action; these are shown by the last column in the following table:
|
|
option |
actual eventuality |
probability of event |
Consequences of event |
Likely consequence |
|
1. |
choose ace |
Ace |
0.077 |
+ £1000 |
+ £77 |
|
2. |
|
not an Ace |
0.923 |
- £1000 |
- £923 |
|
3. |
choose spade |
Spade |
0.25 |
+ £300 |
+ £75 |
|
4. |
|
not a Spade |
0.75 |
- £300 |
- £225 |
If he wishes to minimise his possible losses he chooses a spade (line 4). If, however, he wishes to maximise his gains he chooses an ace (line 1).
[xi] see Conclusion 4 -12 The
best estimate for the rate of misdiagnosis of PVS is of the order of 50%. This
implies that the diagnosis of PVS is essentially a random process.
[xii] whilst diagnosed as PVS.
[xiii] i.e. as in situations where lack of consciousness is accepted as the criterion for withdrawal of ANH; it is argued in Part 3 that ‘lack of consciousness’ is not an appropriate criterion for such decisions.
[xiv] The concepts of ‘a good death’ and ‘personhood’ provide a conceptual structure to enable these decisions to be made (to be discussed in Part 3).